Positive psychology focuses on factors that help promote optimal well-being and investigates the functional consequences of these states [7]. A recent theory by Seligman proposes five pathways to well-being: positive emotions, engagement, relationships, meaning, and achievement [8]. Seligman posits that these pathways are unique predictors of well-being, each of which people pursue and value in their own right.

However, these pathways are linked not only to higher levels of self-reported psychological well-being but also to tangible functional benefits. Experimental and longitudinal studies have found that increased positive emotions lead to social, occupational, and health benefits [9]. Gratitude, for example, promotes both the formation and the upkeep of friendships [10]. Correlational evidence links higher reported focus on engaging and meaningful activities to greater occupational success and higher incomes [11]. Consistent findings suggest that stronger social relationships reduce mortality and help us cope with life stressors [12].

Furthermore, psychological well-being in general, and positive emotion in particular, appear to promote meaning-making in stressful situations, which in turn leads to more effective coping [13]. In a study examining the role of positive emotions in psychological resilience, individuals who experienced positive emotions in the weeks immediately following September 11th were less likely to experience increased depressive symptom levels, controlling for baseline symptoms [14]. Even more interestingly, a recent study showed that increased psychological well-being is a significant protective factor from the occurrence of future episodes of major depression [15]. Indeed, this effect remained even when controlling for baseline levels of depressive symptoms, suggesting that positive mental health is different from merely the absence of psychological distress [16]. Thus, increasing well-being might not only increase positive factors but also reduce negative states and psychopathology.

A variety of researchers and practitioners have translated theories and findings based on positive emotions, well-being, and interpersonal functioning (ie, gratitude, kindness, positive responding, and savoring) into cognitive and behavioral practices deemed positive interventions. The most common outcome measures used to evaluate these interventions reflect aspects of subjective well-being and include increasing cognitive evaluations (such as life satisfaction or happiness), promoting positive emotions, and decreasing negative emotions (or reducing depressive symptoms).

In one of the first empirical investigations of an intervention aimed to increase happiness, Fordyce provided college students with a course that taught participants to mimic the behaviors of happy people [17]. These skills included behavioral recommendations such as increasing physical and social activity and cognitive strategies such as reducing one’s expectations and cultivating a present-focused orientation. This course led to significant improvements in happiness compared with a placebo control condition. More recent research has focused on isolating individual factors and delivering single interventions. A selection of individual exercises with various targets (gratitude, strengths, and reminiscence) are reviewed below.

Thus, various strategies can help promote individual well-being. A recent meta-analysis demonstrated that compared with control conditions, positive interventions lead to reliable and moderate boosts in well-being and improvements in depressive symptoms [23]. Although many of the interventions were brief, single intervention techniques, moderation analysis revealed higher effect sizes associated with “shotgun” approaches, that is, packages that incorporate multiple positive psychology strategies. In light of this, the way to produce the most efficacious intervention might be to combine techniques from previously validated individual exercises into an intervention package.

Positive psychotherapy (PPT) adopted this approach by drawing content from single exercises to create individual and group formats delivered by a single clinician [24]. PPT is a novel and distinct therapy technique that focuses on positive emotions, engagement, and meaning rather than targeting negative states and depressive symptoms. Individual PPT was piloted in a 14-session model for patients with a major depressive disorder recruited from a university counseling center. In a 12-week trial, individual PPT compared with eclectic counseling or eclectic counseling plus medication led to greater improvements in symptoms of depression at the end of treatment [24]. Group PPT was tested in a 6-week, 6-session model for groups of 8 to 11 mild-to-moderately depressed university students. Compared with students assigned to an assessment-only control group, students assigned to group PPT experienced significantly greater symptom reduction and improvements in life satisfaction, and these improvements remained at the 1-year follow-up [24]. These studies indicate the feasibility of these techniques to reduce depressive symptoms in individuals with clinical and subclinical levels of depression. The design of the investigation that we report in this paper translated the individual exercises used in group PPT for online dissemination.

Other programs that have combined content from individual exercises into longer and more involved intervention packages have been delivered in a variety of settings including schools and have also been delivered to military personnel [25,26]. Although outcome studies of these programs are still pending, the reaction from participants has been extremely positive, illustrating the widespread appeal of these techniques. Disseminating these programs, however, is limited by the need to train personnel and the ability of the programs to reach others. A meta-analysis of the school-based Penn Resiliency Program found that only trials delivered by highly trained research team providers produced significant gains [27]. Improving the dissemination of these techniques should be a focus of future research. The study we report here was intended to contribute to this goal.

The brief duration and universal appeal of these strategies make positive psychology interventions ideal candidates for online administration. In fact, one of the first empirical tests of positive psychology interventions was conducted via the Internet using a convenience sample drawn from visitors to the website linked to Seligman’s book Authentic Happiness [28]. The Internet allows dissemination of these techniques throughout the world, anywhere participants have Web access, rather than having to rely on trained professionals to provide access to these techniques. Positive psychology techniques also lend themselves to Internet research, as these techniques might be less stigmatizing than other interventions that deal with difficulties and deficits. In positive psychology exercises, participants talk about the best aspects of themselves, reflect on what is going well in their lives, and reach out to others in their social network in appreciation. The three good things exercise was even integrated into Facebook through an application that provided users with tools to complete the gratitude journaling online and share with friends in their network [29]. Although the application was not evaluated for efficacy, availability through a Facebook application did lead to increased usage. People using the application on Facebook were roughly twice as likely to report good things compared with users of a three good things Facebook group, which is distinct from an application in that it is merely a place where people can post and has no features [29].

Researchers have also used other platforms to promote dissemination of positive psychology interventions. Several individual exercises have been incorporated into Live Happy, a positive psychology app developed for the iPhone. In a naturalistic evaluation, participants who used the app reported increases in their mood and happiness throughout a 2-week period [30]. Participants who completed more activities within the app reported larger boosts in well-being than those who used the app less frequently.

Positive psychology strategies have also been used to augment other interventions to increase usage. In a study of an online body dissatisfaction intervention, participants who completed a gratitude diary were twice as likely to complete the intervention compared with a thought-monitoring and cognitive-restructuring group [31]. This relates to the theory underlying PPT, which contends that focusing on strengths and positive emotions fosters engagement and meaning more so than other techniques. Positive psychology techniques might be more enjoyable and acceptable than other interventions. This is consistent with the overwhelming positive feedback from military and school-based positive psychology programs.

Thus far, we have discussed existing positive psychology approaches and two important limitations of these approaches: feasibility (most interventions are administered in person, often by a therapist) and length (most interventions consist of a single exercise). The goal of the current study was to address both of these issues by offering a Web-based intervention that would last between 2 and 6 weeks. We aimed to systematically examine the effects of increased intervention length on adherence and efficacy, while also gaining basic usability data about the acceptability of an online positive psychology intervention that offers more than a single exercise.

The current study adapted group PPT into an online environment in order to investigate these aims. As previously stated, a meta-analysis of existing positive interventions demonstrated that longer interventions and packages with several techniques typically reported larger effect sizes. Our Web-based trial of positive interventions was designed to investigate these questions by comparing sets of 2, 4, or 6 exercises as well as an assessment-only control condition in terms of usage of the techniques (attrition and self-report) and efficacy (decreases in depressive symptoms). Exercises were drawn from group PPT and included active-constructive responding, a gratitude visit, life summary, three good things, savoring, and strengths. Table 1 provides brief descriptions of each of the exercises.

We posited that similar to the in-person group PPT, the online administration would lead to significantly greater decreases in depressive symptoms. Furthermore, consistent with the findings of the meta-analysis of positive interventions, we predicted that participants receiving more exercises would experience greater benefits.

Participants were recruited through a research portal associated with the University of Pennsylvania’s Positive Psychology Center and Seligman’s book Authentic Happiness [28]. A link available on the websites www.authentichappiness.org and www.ppresearch.sas.upenn.edu invited interested individuals to participate in a research study on positive psychology exercises.

These landing sites offered participants a variety of assessment and research opportunities. From the initial website, individuals could access and receive feedback on questionnaires assessing emotions, engagement, meaning, and life satisfaction. Those who clicked on the link for a research study on positive psychology exercises were informed that they would be randomly assigned to conditions receiving either 2, 4, or 6 positive psychology exercises or an assessment-only control condition over a 6-week period and would be asked to return to the website each week to complete assessment measures. Participants completed consent online by reading the informed consent document and providing an electronic signature.

After completing the baseline questionnaires that assessed demographics as well as mood and depressive symptoms, participants were randomly assigned to one of the four groups. Those assigned to one of the three treatment conditions received instructions for the first positive psychology exercise and were asked to practice that exercise over the course of the following week. Exercises were provided in the following fixed order: three good things, strengths, gratitude visit, savoring, active-constructive responding, and life summary. Participants in the 2-exercise condition, for example, received the first two exercises in this sequence (three good things and strengths) and were encouraged to use these exercises in the following weeks and return to the website to complete follow-up measures. Participants in the assessment-only control condition simply received a message to return the following week to complete additional assessments. Participants in all groups received email reminders each week to return to the site and complete the weekly assessments. After completion of these measures, participants received the next exercise in the sequence if allowed in their study condition.

At baseline, participants completed thorough demographics that included age, gender, ethnicity, marital status, educational attainment, income, and region (first three digits of their zip code). Participants also responded to questions regarding history of psychiatric illness and treatment. During the baseline assessment and at each weekly follow-up, participants also completed the Center for Epidemiological Studies-Depression scale (CES-D), a 20-item self-report scale designed to measure the current level of depressive symptoms [33]. The CES-D contains items assessing negative symptoms such as depressed mood (eg, “I felt depressed”) and appetite disturbance (eg, “I did not feel like eating; my appetite was poor”) as well as positive symptoms (eg, “I felt hopeful about the future”) which are reverse-coded to create a single summary scale of depressive symptoms.

In order to investigate the hypotheses of the current study, we conducted chi-square tests comparing attrition rates across the four conditions. Given the high rates of attrition in Internet-based studies, we conducted analyses of the completers, that is, participants who completed the pretest and posttest measures. We used repeated measures analysis of covariance of the completers to assess changes over the 6-week intervention period in depressive symptoms across the 4 conditions while controlling for baseline levels of symptoms.

Participants included all those who visited the website and consented to participate during a 20-month period from February 2007 through November 2008. Participants were predominantly women (1038/1364, 76.1%) and middle aged with a mean age of 42.38 years (SD 12.08). The largest proportion of participants lived in the United States (743/1364, 54.5%), although many participants were from other countries (621/1364, 45.5%). Most participants were married (602/1364, 44.1%), while 35.0% (477/1364) reported never having been married, and the remaining 20.9% (285/1364) were separated, divorced, or widowed. This was a highly educated sample with 77.2% (1053/1364) of participants reporting they had received a bachelor’s degree or higher. Income was generally high as 24.7% (337/1364) reported earning over US $99,000, and only 11.7% (160/1364) reported earning under US $20,000.

Figure 1 displays the consolidated standards of reporting trials (CONSORT) diagram for the study. Of the 1364 participants who consented to participate, 1140 (83.6%), 966 (70.8%), 851 (62.4%), 770 (56.4%), 705 (51.7%), and 661 (48.4%) completed measures each week at the week 1 through week 6 assessments respectively. At the end of the 6-week assessment period, attrition rates across the active treatment conditions were similar, that is, 55.5% (181/326) for 2 exercises, 55.8% (203/364) for 4 exercises, and 52.7% (168/319) for 6 exercises, but all were significantly greater than the 42.5% (151/355) attrition rate for the control condition (χ² ₃ = 16.40, P < .001).

Compliance (ie, exercise usage) was assessed for each exercise at the 1-week assessment following its assignment and again at each of the subsequent assessments. There were no significant differences between the 6 exercises on use of the techniques. Individuals in the 6-exercise condition were significant more likely than participants in the 4-exercise condition to use both the third (F _1,312 = 5.61, P = .02) and fourth (F _1,313 = 6.03, P = .02) exercises. At the end of the 6 weeks, the three active conditions differed significantly on the number of days that the techniques were used (F _2,358 = 33.84, P < .001). For the 2-exercise condition, mean usage was 6.1 days (SD 3.56); for the 4-exercise condition, mean usage was 7.94 days (SD 4.77); and for the 6-exercise condition, mean usage was 12.1 days (SD 7.36). For 5 of the 6 exercises (all of the 6 with exception of the life summary), enjoyment was related to continued use of the exercise at the 6-week follow-up (r’s = .12 to .39, P’s < .05).

To analyze efficacy of the package of interventions, a repeated measures analysis of covariance was conducted on measures of depressive symptoms. At baseline, the average scores on the CES-D were consistent with mild-to-moderate levels of depressive symptoms (mean 16.93, SD 12.64). There were no significant differences between the groups on baseline levels of depressive symptoms: means were 16.22, 15.95, and 18.10 for participants in the 2-, 4-, and 6-exercise conditions respectively, while the mean for those in the control condition was 17.49. A significant main effect of time (F _1,656 = 94.71, P < .001) demonstrates that depression scores dropped over time in all conditions. Furthermore, a significant condition by time interaction (F _3,656 = 4.77, P = .003) indicates that these decreases differed across the conditions. Post-hoc Tukey HSD (honestly significant difference) tests analyzed the differences among the four conditions. These analyses showed that at the end of the intervention period, participants in both the 2-exercise condition and the 4-exercise condition reported significantly fewer depressive symptoms than those in the 6-exercise condition and numerically (but not significantly) lower than participants in the control condition. Depressive symptoms reported by participants in the 6-exercise condition did not differ significantly from those reported by participants in the control condition. Figure 2 displays mean CES-D scores by condition at each weekly assessment. Estimated marginal means in the CES-D were 12.31, 11.95, 15.89, and 14.53 for the 2-, 4-, and 6-exercise groups and control condition, respectively.

We also analyzed whether use of the exercises played a mediating role in the changes in depressive symptoms. We conducted regressions predicting changes in depressive symptoms while controlling for baseline symptoms investigating both linear and curvilinear relationships with the number of days that exercises were practiced. Neither term was significant, suggesting that the number of days of engagement did not relate to changes in depressive symptoms.